The Graveyard

I built the graveyard before I had anyone to put in it.

If you go to relay1.vc/graveyard right now, you'll see an empty page. I built it that way on purpose, and I'm aware of how strange it is to ship an empty page on launch day. The reason it's empty and the reason it exists are the same reason: most of the companies we co-found won't make it.

That's not pessimism. It's base rates. Most startups fail. Roughly three out of four venture-backed companies don't return capital. The studio model tightens the operating discipline and shortens the runway to signal, but it doesn't repeal the math. We co-found more companies and we co-found them faster, which means our absolute count of failures will be higher than a check-writing fund's, even if our hit rate is comparable. If Relay 1 co-founds thirty companies over the next five years, I expect somewhere between eighteen and twenty-three of them will end up on this page. I'd love to be wrong on the high side. I won't bet against the math to make myself feel better about it.

I'm telling you this on day one because I'd rather you decide whether to work with me knowing the math than have it sneak up on you in year three.

The default is silence

Most studios don't show their graveyard. I get why. There are real reasons. Founders who tried and didn't make it deserve to move on cleanly, and a public obituary doesn't help them get hired into their next thing. LPs reading the next fund's deck don't want a stack of public losses sitting one click away from the wins. And honestly, nobody wants their failures permanently indexed by Google. So the default is silence. The pages get deleted. The websites get redirected to the studio's homepage. The case studies get told only in private rooms over private dinners, and only to people who already wrote the check.

The compounding effect of that silence is what bothers me. When studios hide their losses, the entire ecosystem ends up with a survivorship-biased map of what works. Founders pattern-match on the wins they can see and miss the patterns inside the losses they can't. New studios spin up on glossy thesis decks without any sense of which categories ate everyone alive last cycle. The same mistakes get rerun, slightly remixed, by the next cohort.

I'm not building Relay 1 against any of that. I'm building it differently because of what I believe the public commitment forces in me.

Why pre-commit

Here's what I think happens if you don't pre-commit. The first failure shows up. It hurts. The founder is grieving. You're grieving. There are LP conversations to manage and other portcos watching. The honest thing to do is publish a case study with everything you'd want a future co-founder to know. The easy thing is to let it disappear.

By the time the moment arrives, every incentive is pulling toward easy. I already know the shape of the rationalization. It'll sound something like: "the founder is in a hard spot, let's give them six months before we say anything publicly." Six months will become a year. The case study will get drafted, then shelved for the next raise. Then the next one.

I built the page before any of that has a chance to happen, because the only way I know how to commit to something hard is to build the thing I'd have to take down to break my word. The empty page is a tripwire. If a year from now you visit this URL and find a 404 or a redirect, you'll know exactly what happened, and so will I.

What goes on the page

When a Relay 1 venture gets sunset, the entry covers four things.

The original thesis. What we set out to build, for whom, and why we believed it was a category worth co-founding into. Not the pitch deck version. The honest version, including the doubts we had at formation and the things we wanted to be true that we couldn't yet prove.

The decision points. The specific moments where we doubled down, where we pivoted, where the signal said wind down sooner than we acted on it. Studios are made of these moments more than they're made of strategy. I want a future operator to be able to read down the timeline and watch the calls happen in real time, including the ones we got wrong.

What tipped it. The specific signal, market shift, capital outcome, or product reality that ended the run. Not a vague "market conditions changed." The actual thing. The week it became clear. The conversation that made it undeniable.

What changed at Relay 1. How the studio's playbook updated as a result. Which screens we added at the next intake. Which formation patterns we won't repeat. Which assumptions about a category we retired. The whole point of running a studio is that the operating system gets sharper with each venture. The graveyard is where that sharpening shows its work.

One test I'll hold the case studies to: if a sentence in the post-mortem could have appeared unchanged in the original pitch deck, it shouldn't be in the post-mortem. The deck was the story we wanted to tell. The case study is the story that actually happened.

The case study standard

Every entry links to a real case study, and by case study I mean a long document, three thousand words minimum, with the founder's own voice in the parts they want to write themselves. The founder owns the narrative on their own work. They get to write the story of what they built and why it ended, in the words they want to use. Relay 1 contributes the studio's perspective, the market context, and the operating system lessons.

There are guardrails I care about. The case study is a learning artifact, not a tribunal. We don't name vendors who failed us, customers who didn't sign, or specific employees who didn't perform. We talk about decisions and patterns and structural realities, because those are the things a future operator can actually use. The founder gets the final read on anything attributed to them, including a veto on details they don't want public. The studio's section is mine to write, and I'll write it whether the founder likes the conclusions or not, but the founder always sees it before it ships.

If a founder absolutely doesn't want a public case study, they can opt out. The entry still goes on the page with the thesis, the dates, and a one-line summary. The detailed case study stays private. I'd rather that than coerce a story out of someone who isn't ready to tell it.

The operational specifics

Specifics, so this is a commitment and not a vibe.

Sunset means a Relay 1 venture has filed dissolution, completed an asset sale, or stopped operating with no intent to resume. Quiet pivots and pauses don't count. The clock starts the day the wind-down decision is made by the founder and the board.

The case study ships within ninety days of that date. The founder gets the first thirty to draft their section. I get the same thirty to draft mine. The last thirty are for exchanging reads, arguing about what stays in, and getting it across the line. If a founder is in active legal proceedings related to the wind-down, the ninety-day clock pauses until those are resolved, and that pause itself gets noted on the page when the entry goes up.

The studio's section is written by me, not delegated. The founder gets veto power over anything attributed to them and over any detail about the company they don't want public. They do not get veto power over my read on what the studio got wrong.

The page stays up indefinitely. Entries don't get archived, hidden, or paywalled when the studio raises a new fund or rebrands. If Relay 1 ever takes this page down, that itself is a signal worth more than anything I could write on it.

The trade

This is the trade I'm asking founders we co-found with to make. You'll get the studio in the work with you, weekly, from formation. We'll ship the build, contribute the brand, hold the operating cadence, and stay through scale. We'll write you onto the cap table as a real co-founder, not a hired CEO. In exchange, when the venture winds down, the case study goes public. Your call to wind it down, not mine. Your name attached. Your decisions visible. The full story.

I think that trade is worth it for two reasons.

The first is selfish. It forces me to be a better operator. If I know every wrong call I make is going to live on a public page with my name on it, I make fewer wrong calls. I name the hard read sooner when the data says name it. I push back harder when a founder is in love with a wedge that the market keeps refusing, even when pushing back is the least comfortable thing in the room. Pre-commitment is the cheapest form of discipline I know.

The second is for the founder I haven't met yet. A few years from now, somebody is going to read this page before they take a meeting with me. They're going to see a stack of sunset case studies. They're going to read what went wrong, what I missed, what I owned. They're going to read what didn't go wrong too, in the case studies of the ventures that worked. Then they're going to decide whether they want to co-found their company with the operator who ran those plays. I'd rather they have the full ledger than the highlight reel.

What this changes about how I operate

About eighteen months in, a venture starts sending the kind of mixed signal that's worse than a clean miss. A handful of customers are obsessed. The next tier is polite. The metrics aren't falling, but they're not compounding the way they need to be by now. The founder is running on fumes and still believes, and you can see that they need someone to either confirm the belief or push back on it. This is the moment most studios lose discipline. The capital is already deployed, the team is built, the brand is in market. Killing it feels like admitting you were wrong, and continuing it feels like loyalty. So the studio continues, quietly, for another year, hoping the signal turns.

I want the graveyard to be the thing I think about in that exact moment. Not as a threat. As a clarifier. The founder owns the majority of the company and the call to keep going or wind down is theirs to make. My job is to bring the question into the room earlier and more honestly than studios usually do. The question stops being "do I want to admit I was wrong" and starts being "what would the case study say if we shipped it next quarter." If the answer is "we kept going past the point where the signal told us to stop, because we were attached," that's the case study neither of us would want to publish. So I push the conversation forward instead of waiting for it to arrive on its own.

That's the operating system this page is supposed to enforce. Not a guarantee that every venture works. A guarantee that when one doesn't, we knew it sooner, killed it cleaner, and learned from it more honestly than we would have if no one were watching.

I don't think of the graveyard as a confession. I think of it as a contract, written in advance, with the version of me who'll be tempted to look the other way.

When the first one happens, it'll show up here.